As it happens, how many doctors a country has working per capita is a powerful factor. We prefer to use the reciprocal - how many patients per doctor. Either measure is a good way to measure how effective an educational system is
and how well a government invests in medical infrastructure. Of course, there are many complications: are individual citizens near to a doctor; can they afford to go; is there an epidemic raging? Then there are issues with if the doctor any good and can he or she treat you - a podiatrist, whoever skillful, is not much use during a difficult pregnancy.
We would think every government on the planet would find time to ask during a census 'are you a doctor?' and 'are you actively working?'. Better, of course, is to measure at least every year.
It turns out the responses are literally all over the map. So one gets a sparse matrix of countries (rows) and years (columns). For our purposes, we'd like to use the most recent "count" of doctors. But, how are we to know the number has any reliability?
First the raw data - just the number of doctors reported: